Offermic

Mercor AI Interview: Format, Retakes, Rules and How to Prepare

How the Mercor AI interview works per Mercor's own docs: length, format, retakes, practice interview, AI-use rules, setup, and how to prepare your answers.

Updated

Written by

The short answer

The Mercor AI interview is a roughly 20-minute video and audio conversation with an AI interviewer that asks role-specific questions, often about your past experience. Mercor offers an unlimited free practice interview and up to three official retakes. Prepare detailed stories about your work, use Chrome with a camera and microphone, and answer in your own words.

Key takeaways

  • Most Mercor AI interviews take about 20 minutes, according to Mercor.
  • You can take a free Practice Interview as often as you like; it does not affect opportunities.
  • Official interviews can be retaken up to three times across applications for that interview.
  • Completed interviews carry over to other applications that need the same interview.
  • Use a computer with Chrome (recommended), Edge or Safari. Firefox and mobile browsers are not supported.
  • Mercor's rules ask you not to use LLMs to compose your answers.

The Mercor AI interview is a video and audio conversation with an AI interviewer that asks role-specific questions, records a transcript and evaluates your performance for the listing. Mercor says most interviews take about 20 minutes. You can take a free Practice Interview as many times as you like, and retake the official interview up to three times. To prepare, get ready to discuss your past work in depth, test your setup in Chrome, and answer in your own words.

Everything about Mercor on this page comes from Mercor’s public talent documentation and website, checked on 2026-09-23. Offermic is not affiliated with Mercor.

What is Mercor?

Mercor is a platform that connects professionals with AI companies and enterprises. Its homepage describes the company as “organizing human intelligence to power the AI economy,” and lists expert work for professionals such as physicians, lawyers, engineers and consultants. Many roles involve helping AI labs train and evaluate models using professional expertise.

Because Mercor matches experts at scale, the AI interview is central to its process. Mercor’s FAQ says: “Our AI interview is the standard and only method for screening candidates,” and that interviews cannot be done by chat, call or another medium instead.

How does the Mercor AI interview work?

From Mercor’s candidate FAQ:

  • Who asks: An AI interviewer poses role-specific questions and generates a transcript of your responses.
  • Length: Most interviews take approximately 20 minutes; some are shorter or longer.
  • Content: “Some interviews are technical and require deep discussion of prior experience; others test language or reasoning skills.”
  • Language: Interviews are in English unless otherwise specified. Captions are on by default.
  • Extras: Some interviews require full-screen sharing, and some include a whiteboard that saves automatically.
  • Reuse: “Each unique interview is only necessary to complete once.” If another application needs the same interview, your result carries over.
  • Starting: You start the interview from the application’s step list on your dashboard, and you don’t have to do it immediately.

The interview sits alongside Mercor’s assessments, which its docs describe as structured evaluations of your domain expertise, problem solving, ability to follow instructions and “the quality and clarity of your written or verbal responses.” Mercor notes that even when two listings use the same assessment, “the scoring and weighting may vary.”

Retakes and the practice interview

Practice Interview Official AI interview
Attempts Unlimited Up to 3 retakes across applications for that interview
Affects opportunities No Yes
Shared with companies No Used to evaluate your candidacy and recommend your profile
Tied to an application No Yes

To retake, Mercor’s docs say to open the Assessments tab on your dashboard, find the submitted interview, click the three-dot menu and select “Retake interview.” If you’ve used all attempts and had a genuine technical problem, you can email support with a screenshot or screen recording to request a one-time retake, but approval isn’t guaranteed.

Tip: Do the Practice Interview at least once before your official attempt. It’s the closest thing to the real format, and it costs you nothing.

Setup requirements

Mercor’s documentation lists these requirements:

  • Browser: Chrome (latest) recommended; Edge and Safari supported; Firefox and mobile browsers not supported.
  • Device: A computer with a working camera and microphone. Test both in the waiting room.
  • Visibility: Your full face visible and your voice clear throughout.
  • Environment: Quiet, distraction-free and with a stable connection.
  • Audio: Headphones and an external microphone are recommended.
  • Materials: Pen and paper, plus a calculator if relevant.
  • Troubleshooting: If the bot doesn’t respond, wait 30 to 60 seconds, refresh, and try disabling VPNs and browser extensions.

The “Start Interview” button stays grayed out until camera, microphone and (if required) screen-share checks pass.

What does the Mercor AI interviewer evaluate?

Mercor says the AI evaluates “your performance for the listing,” and that its goal is consistent evaluation across applicants. It doesn’t publish its scoring rubric. Based on what Mercor says the interviews contain, expect to be assessed on:

  • Depth of real experience. Can you explain what you did, how and why, in detail?
  • Domain knowledge. For technical and professional roles, whether you know your field.
  • Reasoning and communication. Whether your answers are structured and clear.

In other words, the interview rewards people who can talk concretely about work they’ve actually done.

How should you answer Mercor interview questions?

Prepare for deep dives on your experience

Pick three or four projects from your resume that fit the listing. For each, be ready to explain:

  • The problem and why it mattered.
  • Your specific role and decisions.
  • The trade-offs you considered.
  • The outcome, and how you measured it.
  • What you’d do differently.

This is the STAR method with extra depth. MIT’s career office suggests putting most of an answer into the Action: what you personally did.

Example answer — senior clinician applying for a medical expert role, “Tell me about a complex case where your judgment made a difference.”

  1. Situation

    On a busy night shift, a patient admitted for a routine infection had lab results that were within normal ranges individually but trending in a worrying direction together.

  2. Task

    As the senior clinician on shift, I had to decide whether to escalate care before the patient met any formal criteria.

  3. Action

    I reviewed the trend across the last three sets of results rather than the latest values alone, re-examined the patient, and discussed the picture with the on-call specialist, explaining exactly which changes concerned me. We agreed on closer monitoring and an earlier review, and I documented my reasoning so the day team could follow it.

  4. Result

    The patient's condition changed early the next morning and the team was already prepared, so treatment started without delay. Afterwards I shared the case at our team meeting as an example of reading trends, not single values.

This example is written for illustration; it isn’t a real case.

Keep answers concise and avoid long silences

Mercor’s FAQ says to articulate “clearly and succinctly” and avoid prolonged pauses, since extended silences can make the AI interject. If you need a moment, say “Let me think for a second.” If it interrupts before you’re done, Mercor says you can let it know you weren’t finished.

Ask for clarification

If a question is unclear, ask the AI to repeat or clarify it. Answering the wrong question wastes your time.

Explain complex ideas simply

Many Mercor roles involve explaining expert reasoning. Practice explaining a technical concept from your field to an intelligent non-expert. See explaining something complex.

Weak answer

“I worked on a lot of data projects at my last company, mostly in Python, and I did analysis and some machine learning too. It went well and the team was happy with the results.”

Why it falls flat: Generic and unverifiable. An AI or human interviewer can't evaluate skills that aren't described.

Strong answer

“The project I'd highlight is a churn model for our subscription product. I owned feature engineering and evaluation. The first model looked great on paper but leaked future information through a billing field, so I rebuilt the training set with strict time cut-offs. The honest model was less accurate on paper but held up in production, and the retention team used it to target outreach.”

Why it works: Specific project, clear ownership, a real technical decision with its trade-off and an outcome. It invites a good follow-up.

What kinds of questions should you expect?

Mercor doesn’t publish its question lists, and it asks candidates not to share specific questions. Based on how Mercor describes its interviews (technical deep dives on prior experience, or tests of language and reasoning), you can prepare for these broad types:

  • Background: “Walk me through your experience” or Tell me about yourself. Keep it to about 90 seconds and connect it to the listing.
  • Project deep dives: “Tell me about a project where…” followed by questions on your decisions. Prepare three or four projects in depth.
  • Problem solving: “Walk me through how you solved…” See complex problem you solved.
  • Domain knowledge: Questions that check you know your field, such as clinical reasoning for medical roles, legal analysis for legal roles or code and system reasoning for engineering roles.
  • Explaining your reasoning: Many expert roles involve evaluating AI output, so clear, step-by-step explanations matter.
  • Achievements: “What are you proudest of?” See proudest achievement.

For each type, the winning pattern is the same: a specific example, your own decisions, and a clear outcome.

Rules on using AI during the interview

Mercor’s FAQ asks candidates to follow its LLM Usage Policy: you can do grammar and wording checks where relevant, but should “refrain from utilizing large language models to compose your responses, rate model responses, generate rationales or explanations, or evaluate code.” Mercor also asks candidates not to share specific interview questions or screenshots online.

Practically, answers written by an AI also tend to be generic, which is the opposite of what an experience-focused interview rewards. Use AI tools to practice before the interview, not during it.

Privacy: what Mercor says it does with your interview

According to Mercor’s FAQ:

  • It retains video/audio recordings and an AI-generated transcript of official interviews to evaluate your candidacy and recommend your profile to employers.
  • It does not sell your interview data.
  • Your interview data is not used to train AI models.
  • Practice interviews are not shared with companies.
  • It can’t share recordings or transcripts with you.

A 3-day Mercor prep plan

Day 1: Re-read the listing and your resume. Pick three or four projects and write STAR notes with extra depth on decisions and trade-offs.

Day 2: Practice out loud. Explain each project in two minutes, then answer “why did you choose that?” and “what would you change?” for each. An AI mock interview helps here: Offermic, the app we make, runs spoken mock interviews based on your resume and a job posting, asks follow-ups, and scores each answer 0–100 with feedback. Practicing with a colleague works too.

Day 3: Take Mercor’s Practice Interview in Chrome with your real setup. Fix anything that went wrong (audio, lighting, pauses), then take the official interview when you’re fresh.

After the interview

Mercor says it usually takes 2 to 4 weeks to hear whether you’ve moved to the next stage, because of application volume, and encourages candidates to keep applying to other roles. Because interview results carry over, one strong interview can support several applications.

Questions people also ask

How long is the Mercor AI interview?

Mercor says most AI interviews take approximately 20 minutes, although some are shorter or longer depending on the role.

Can I retake the Mercor AI interview?

Yes. Mercor allows up to three retakes of an interview across all applications that use it, via the Assessments tab on your dashboard. If attempts run out because of a technical problem, you can ask support for a one-time retake with evidence, but approval is not guaranteed.

Is there a practice interview on Mercor?

Yes. Mercor offers a free Practice Interview that is not tied to any application, is not shared with companies and does not affect future opportunities. You can take it as many times as you like.

Can I use ChatGPT during the Mercor interview?

Mercor asks candidates to follow its LLM usage policy. You may do grammar and wording checks where relevant, but should not use large language models to compose responses, rate model responses, generate rationales or evaluate code.

How long does Mercor take to respond after the interview?

Mercor says it usually takes 2 to 4 weeks to hear whether you have moved to the next stage, because of application volume. You can keep applying to other roles in the meantime.

Sources

  1. AI Interview (candidate FAQ), Mercor (2026-09-23)
  2. Assessments, Mercor (2026-09-23)
  3. Mercor homepage, Mercor (2026-09-23)
  4. The STAR Method for Behavioral Interviews, MIT Career Advising & Professional Development (2026-09-23)

Written by, Founder

Can Garip is the founder and developer of this app. He builds the AI mock interview product and writes its interview preparation guides.

Drafted with AI assistance, then edited and fact-checked by the author.

How we write